AC Filtering, Grounding Boxes, Linear PSU and Balanced Power.

[YashN]

Thanks. Sreamer as I use might respond differently. I will check it out when I have got the time.

 

A couple of ways you can potentially check:

 

1. The computer's list of USB devices detected. On macOS, I see this in two places: the System Properties -> Sound -> Output Tab. I also see it in the Utilities -> Audio MIDI Setup -> Audio Devices list.

 

2. Your audio player's Audio device list. In Audirvana+, I see that Active Device capabilities are grayed out when it's disconnected and when the handshake is completed, the resolution/rates light up green again.

 

There must be similar ways to see how fast that handshake completes in other OSes (Win and Linux). For a streamer I don't know how you could check this, but in principle, there's no reason for the handshake to take longer or much longer.

 

I have a third way of knowing the handshake is completed in my setup, and that is a couple of very low volume 'clicks' that the DAC does. This, however, may not occur with other DACs.

[YashN]

LOD batteries.... John mentioned regulation... I use the Phone bank I never thought of its voltage out and attached regulators.....

Any DIY regulators that can help the out DC.....

It would be interesting to know how best to implement a battery bank, capacitor + regulator or more.

 

Hey Middy, any reason for the over-use of full-stops in your posts?

[YashN]

The Entreq drain wire implementation is intriguing but could be interesting to try.

 

Must check the implementation of drain and shield and their effects on signal or noise if any. I think it could be a paper by Jim Brown showing measurements of those.

 

Had it open in a tab but we referenced it way earlier in the thread, and a concept mentioned by speedskater: SCIN: Shield Current Induced Noise - Jim Brown.

[Middy]

Yes normally I only use one hand on my mobile...... is a bad habit as a pause in conversation. My thumb just hits nice on the full stop.

If it's driving people mad I'll stop.

 

Apologies if it has...... ?

The main worry was 3.7 to 5V and noise. I thought the battery bank would be really clean but noisy regulators? Is this another issue we can diy fix?

Get that extra bit out of the DAC?

 

Cheers

And Sorry YashN

 

 

 

Sent from my SM-G900F using Computer Audiophile mobile app

[YashN]

Let's discuss this point. some form of ground connection is always needed for the return current to flow. Without ground connection USB connection cannot function. We are discussing the possibility to lead that return current through other path than the USB cable itself.

What's the difference between 'perfect star ground' and my case?

 

In some cases, I've even seen that a return could flow in the other cable of a twisted pair - ouch! not sure this is what we want though.

 

This long text will provides a lot more clues for overall grounding and isolation. It's long but it is really well-made and illustrated with various examples:

 

Audio Component Grounding and Interconnection - diyAudio

[YashN]

The main worry was 3.7 to 5V and noise. I thought the battery bank would be really clean but noisy regulators? Is this another issue we can diy fix?

Get that extra bit out of the DAC?

 

Thinking of the DAC here (or sometimes other devices), I think we need to check whether the device has its own internal regulators, and provide it with enough voltage to remain in regulation mode.

 

Additionally, what I am asking myself is whether straight connection of batteries (and let's say a capacitor) to the device is enough, or we should think of pre-regulating this.

[Cornan]

Thinking of the DAC here (or sometimes other devices), I think we need to check whether the device has its own internal regulators, and provide it with enough voltage to remain in regulation mode.

 

Additionally, what I am asking myself is whether straight connection of batteries (and let's say a capacitor) to the device is enough, or we should think of pre-regulating this.

 

Interesting thoughts! Do you mean something like the JKenny Regen (LifePO4) tweak for a DAC?

 

 

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[Cornan]

Here is something interesting http://electronics.stackexchange.com/questions/55445/does-a-self-powered-usb-device-have-to-connect-to-vcc-from-the-host-port

 

Especially this part. Quote: Mind you that levels for D+ and D- are rated for +3.6V max, not to 5V as you might expect!

 

 

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[Superdad]

Yes, it is true that off the shelf USB cables usually have the shield connected to ground at both ends...

 

Um, sorry Cornan, but that just is not true. All standard USB-spec cables have the shield/shells connected at both ends and the pin 4 ground wire is separate and not tied to those shields.

However, the shields of the USB A and B jacks at the host and device are soldered to the ground plane, and 99.5% of the time pin 4 is also connected to the ground plane.

[Cornan]

Um, sorry Cornan, but that just is not true. All standard USB-spec cables have the shield/shells connected at both ends and the pin 4 ground wire is separate and not tied to those shields.

However, the shields of the USB A and B jacks at the host and device are soldered to the ground plane, and 99.5% of the time pin 4 is also connected to the ground plane.

Thanks Alex! I know, but it is just more simple to explain it that way since the barrels is connected to ground. BTW, do you know if the barrels have to be made of metals or if can they be made of for example plastic if you are using unshielded wires?

 

In lack of electrical tape my own experience covering the USB barrels was postponed! ?

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[Cornan]

I was going to test the different combos today, but the connections were a bit fiddly, so I removed my iMac-to-SS amp chassis-to-chassis connection, connected one side of the jack the shiel (computer-side) and also both sides to the shield, and now the DAC isn't recognised.

 

So, it looks like this is one leg of the very circuitous path for a ground loop, especially since I removed my AC Filter experiments.

 

I've disconnected everything again and put the DIY cable back in the workshop to re-work it. I managed to make it quite short, but it becomes difficult to reach those two jacks + shield to make the connection, so switches with some wires will probably be used in the next iteration.

 

Additionally, I'll probably need a 3-way toggle: one with GND connected for handshake, one with an isolated GND, and one with the GND completely disconnected (for post handshake).

 

Getting things to work in a robust way without any fiddling and open wires and no circuitous route looks like a much-needed better way of testing various combinations.

 

I'm taking the circuitous route as similar to the GND just cut off, which iFi says isn't recommended. Actually they said the GND has to be lifted in a safe way and in accordance with the USB standard, so that's what I'll be looking at.

 

I figure once this is made correctly, nothing prevents the chassis-to-chassis grounding to be re-implemented properly.

 

Now, here is why this is quite circuitous as I found today:

 

- My iMac has a 3-prong power cord, but my SS amp doesn't...

 

- Chassis link was from the iMac to the SS amp

 

- However, there was no chassis link between the iMac and the Tube Amp

 

- The Tube amp isn't connected to the SS amp and the Tube amp has a 3-prong power cord

 

So, I am not sure how the SS chassis link to the iMac is actually causing the handshake to occur or not occur when it's removed. Sounds odd to me like that.

 

Here is the picture that Karin at Sbooster showed r_w some time ago at the VBus Isolator thread:

 

 

Remember that any ICs with ground and/or shield in contact with ground, chassi-to-chassi grounding and mains AC cables with safety ground connected to a single power strip will serve as a "handshake path" from PC to DAC. Ground currents will take any low impedance path available for the task.

 

The way I see it you want to limit the ground currents to the path you choose for them...by closing the unneccessary paths to ground with unshielded ICs and battery supplies..and make the safety ground as good as possible. YMMV ofcourse.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[YashN]

All standard USB-spec cables have the shield/shells connected at both ends and the pin 4 ground wire is separate and not tied to those shields.

However, the shields of the USB A and B jacks at the host and device are soldered to the ground plane, and 99.5% of the time pin 4 is also connected to the ground plane.

Hey Alex, do you have any thoughts about testing various combinations of disconnecting the jacks and the shield at each end for SQ?

 

This would be in a normal case, of course, not including a Regen in the chain, just the computer straight to the DAC through USB.

[YashN]

Interesting thoughts! Do you mean something like the JKenny Regen (LifePO4) tweak for a DAC?

 

I know he did some battery-to-clock and/or battery-to-USB receiver implementation, perhaps he went further and did a full battery power.

 

What I'm wondering is whether I need to regulate the battery power to the DAC, and if so, do I need to pay attention to any internal regulators.

 

In the latter case, I'd have to ensure I have enough battery voltage for the powered circuit to remain within its internal regulated mode.

 

If we don't do this, we may think the pre-regulator we added is causing 'noise' when the actual systemic effect is lack of internal regulation...

[YashN]

Does anyone know if there's a sort of handshake when the type/format of the track changes? That might explain what Cornan was saying.

[Cornan]

Does anyone know if there's a sort of handshake when the type/format of the track changes? That might explain what Cornan was saying.

Handshake Packet

Handshake packets report the status of a data transaction and can return values indicating the reception or

rejection of data, halt conditions, or flow control. They are built only by PID.

There are three possible handshake packets:

• ACK — acknowledgment that the packet was successfully received

• NAK (Not Acknowledgment) — temporarily unable to send/receive data; there is no data to send

concerning the interrupt transfer

• STALL — the endpoint is halted or a control pipe request is not supported, intervention from the

host is required.

A handshake packet is issued or checked by hardware. If data is received correctly, the USB module issues

an ACK. If the endpoint is busy and is not able to receive a data packet, or has nothing to send, hardware

issues a NAK. If you halt the endpoint, hardware issues a STALL. If there is token or data error, hardware

does not issue a response.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[YashN]

Handshake Packet...

How does that help in determining whether there is a handshake on track format change?

[Cornan]

How does that help in determining whether there is a handshake on track format change?

I am just posting my finding regarding the handshake. As I mentioned before the way I personally know that the handshake have been done is when the music starts. Since I cannot find another way to know it on my streamer. The actual handshake is surely done quicker...but in exactly what instance I do not know and cannot tell with Aries Mini.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[YashN]

I am just posting my finding regarding the handshake. As I mentioned before the way I personally know that the handshake have been done is when the music starts. Since I cannot find another way to know it on my streamer. The actual handshake is surely done quicker...but in exactly what instance I do not know and cannot tell with Aries Mini.

 

Wasn't what I was asking though.

 

In your setup when GND is disconnected, and a track in Format A (let's say DSD) is playing, can you successfully change to a track of format B (PCM)?

[Cornan]

Wasn't what I was asking though.

 

In your setup when GND is disconnected, and a track in Format A (let's say DSD) is playing, can you successfully change to a track of format B (PCM)?

I would'nt know since I use Tidal streaming as my one and only source. This might be something that @r_w knows about though...

What I do know is that when the handshake is active it is activated until you power off the DAC. I would be surprised if you could'nt change between different formats when the handshake is activated, but cannot say for sure.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[r_w]

Handshake stays alive regardless of format or sample rate change... it's not about the data (music play) it's about the two units being initially introduced... once introduced and handshake occurs the job is done until either the USB cable is disconnected or either unit is powered down.

 

 

 

I would'nt know since I use Tidal streaming as my one and only source. This might be something that @r_w knows about though...

What I do know is that when the handshake is active it is activated until you power off the DAC. I would be surprised if you could'nt change between different formats when the handshake is activated, but cannot say for sure.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[YashN]

Handshake stays alive regardless of format or sample rate change... it's not about the data (music play) it's about the two units being initially introduced... once introduced and handshake occurs the job is done until either the USB cable is disconnected or either unit is powered down.

That's what I thought as well, but do you know what communication if any occurs between the DAC and the computer on changing formats?

 

During my tests today, as long as I stayed on the same format with one jack-shield connected the DAC would play, but if I changed format, it wouldn't play the new track.

 

I would then have to power-cycle the DAC.

 

I have removed the test rig for now, but one configuration provided a mind-boggingly good SQ, but it's rather unusable because of the format change. That was with the DAC-end shield-jack connection ON and the Computer shield-jack OFF and with the Filter ON. Now, all this to be taken with a big grain of salt as firstly, I had inverted the switch position semantics in my mind, and secondly because of that format change issue.

 

If I managed to get the DAC working with both shield-jack disconnected, then a GND loop is still extant, though how, no idea. Only the iMac and the SET Tube Amp are 3-prong powered. The DAC is powered by the iPhone charger and that's 2-prong.

 

So, I'm now checking out how to perform a ground isolation, but it looks like I'll never be able to really totally disconnect it for long because of the format change issue.

[YashN]

All right, so here's what I have:

 

- This morning: Filter on (computer side), No GND Line, Jack-shield connection supposedly ON on DAC side, and supposedly OFF on Computer side. The big issue I had was when putting the jack-shield ON on the DAC side after 'handshake', I wasn't able to change track if the format was different.

 

This doesn't seem like a good way of architecting noise-removal though.

 

So, I figured I'd have to re-connect GND (sigh), but at least get this filtered somehow.

 

- Tonight: So tonight, I re-worked the cable, removed the filter + termination circuit and built a little Ground-Lift circuit. This circuit is between the computer GND and the DAC GND (unfortunately). However... It does look like the much, much reduced noise is still there. In this configuration, both jack-shields are OFF! (If I disconnect it DAC-side, the sound becomes choppy - this doesn't sound good at all, not just for SQ, but it's scary to think what that could do to the DAC. If I connect the computer-side, the DAC goes blank).

 

Now, technically, the computer-side of the GND connection before this circuit on the way to the DAC should go to a Star-Ground. I had one when I was building the AC Filter and testing it but for now it has been removed, so currently, there's just the GND Lift circuit that is in between the computer and the DAC.

 

The impression I get is that of greater power/volume/clarity simultaneously. And beware: if a record has some hiss in the background, even this is revealed. On OP8 - Leather on the DALI CD, the initial recording has some background, until he sings and then the hiss goes down quite a lot. With a normal cable, it's not brought to your attention. With this new cable and circuits, it is.

 

A lot of micro-detail is revealed on top of the impression of my DAC having more power. I actually had to check the Volume position of the DAC, and it's just like I am used to, except it plays louder.

 

I suppose these are what we stand to lose with dirty power and a normal cable: losses in detail and in overall dynamics translating in overall power loss.

 

So, I could have been disappointed to have to re-connect that GND between the two, but it appears I get a lot of the benefits, much as before.

 

In addition, I can swap track format without having to connect or re-connect anything.

 

Now, I also hesitated to hook in my Computer-side filter circuit as well, but after a couple of close listening sessions with Sting - Sacred Love on DSD as well as some tracks from Sade's Best Of, I just did.

 

Result: More Power and micro-detail to us.

 

And of course, let's not forget the initial goal of this: just to remove the 'digititis' or 'glare' or 'hash'. Haven't heard it in a while.

 

It's rather incredible, and without exaggeration, it sounds like I have a new DAC.

 

One has to mull over all the losses we expose our setup and SQ to if we just connect our computer to our DAC with a normal USB cable.

 

And to think I haven't even worked on the Linear-Regulated PSU (LPR for short henceforth)...

 

Madness I tell you, madness!

[YashN]

I am currently listening to Depeche Mode's "Two Minute Warning" and hearing it like I never have before.

 

[YashN]

Currently playing: Liane Foly's acoustic version of André Manoukian's "Doucement".

 

The sound is powerful, detailed, and no trace as far as I can tell of glare. This is very addictive! And I'm not even a big fan of Liane Foly - this is my gf's album.

 

9pm and I haven't eaten.

[Cornan]

Great news YashN! I know the feeling! I think you understand now why it is impossible to go back.

 

Sent from my Nexus 7 using Computer Audiophile mobile app